Re: Integer moves for fp regs

[Bill Currie <bcurrie at tssc dot co dot nz>]

Michael Hayes wrote:
> 
> md.texi has this tantalising comment to say:
> 
> "There may also be a need to support fixed point @samp{mov@var{m}}
>  instructions in and out of floating point registers.  Unfortunately, I
>  have forgotten why this was so, and I don't know whether it is still
>  true."
> 
> Does anyone know the current situation regarding this?
> 
> I see that caller-save expects that it can save/restore hard floating
> point registers with integer move instructions, provided the registers
> also support integer modes.  Unfortunately, this is not true for the
> C4x and I'm wondering what other parts of GCC make this assumption.

The i860 uses the fp unit to do integer multiplies (fmlow.dd).  Now,
this might be a moot point with the difficulty of getting an i860, but I
wouldn't be surprised if other CPUs do this.  Hmm, doesn't the ix87
support integer multiplies (mind you, who actually uses that feature
when the ix86 has integer mult in the first place?

Bill
-- 
Leave others their otherness